Abubakar Abid on AI for Genomics, Gradio, and the Fatima Fellowship
The Gradient: Perspectives on AI
00:00
PCA, Principal Components Analysis
In biology and medicine, particularly, it's oftentimes used to remove noise in your data. But one of the things that we noticed is that PCA doesn't necessarily remove the noise. It just removes anything that isn't dominant signal. So what we developed was this thing called contrastive PCA. And so if you do that, you end up getting representations which are usually have very little noise, which are very rich in signal.
Transcript
Play full episode
